Most studies of the association of suicide and suicide attempts with "depression" have methodologic limitations, e.g., no structured diagnostic interviews, restricted to one or other domain of potential predictor variables or chart reviews to quantify suicidal behavior. We have implemented the most comprehensive, psychobiological, prospective study of predictors of suicidal behavior in major depression. The study utilizes a unique cohort of patients hospitalized with a Major Depressive Episode comprising two demographically comparable groups of past attempters and nonattempters) who receive extensive clinical and biological baseline assessments as part of the MH62185 Center for the Neuroscience of Mental Disorders at NYSPI/Columbia University. We are now requesting funding to continue the follow-up component of this ongoing prospective study. Patients are reassessed for clinical state, life events, treatment and suicidal behavior at 3, 12 and 24 months after discharge. Determinants of suicidal behavior are derived from multiple domains including psychopathologic, genetic, biologic, psychologic, familial and psychosocial. Thus, development of a general explanatory or predictive model of suicide attempt behavior requires assessment af potential risk factors from multiple domains in the same patient population in a prospective follow-up study. We propose a stress-diathesis model which posits that risk factors are required from both the stress domain and from the diathesis domain in suicide attempters. The major hypotheses of the present study are that when a major depression is present, other factors, particularly from the diathesis domain are paramount, e.g., characterological traits (such as aggression/ impulsivity), a vulnerability to pessimism (more suicidal ideation, fewer perceived reasons for living, more subjective depression) and biological measures (regional brain serotonin receptor changes, HPA dysfunction, serotonin-related gene markers) to distinguish attempters from nonattempters, both retrospectively and prospectively. Thus far, we have studied 389 patients at baseline with follow-up data on approximately 85%. In the next 5 years we anticipate entering another 200 subjects. Baseline factors distinguishing pastattempters and nonattempters include aggression/impulsivity traits, pessimism, low serotonergic activity, family history of suicidal acts, and HPA axis dysfunction. Future suicide attempts are related to past suicide attempt history, pessimism, aggression/impulsivity and low serotonergic activity, all correlates of past attempts. An expanded study population will permit testing of newly identified potential predictors found to be correlates of past suicidal behavior (abuse history, birth difficulties, early separation, head injury, neuropsychological testing, HPA axis indices and PET quantification of the serotonin transporter and 5-HT1A receptor). We will also replicate and integrate previous predictors with newly confirmed predictors.